Inversion Of Seismic Prospecting Parameters Using Improved Genetic Algorithms

Seismic inversion plays a key role in geophysical prospecting. As most of geophysical inversions appear as objective function optimization with multi-extreme, It is the difficulty that we apply the nonlinear inversion theory to seismic inversion. This study has considerable academic value as well as wide prospects.This paper introduces the concept and theory relevant to the study. It deeply researches and expounds Genetic Algorithms and Simulated Annealing. With fully absorbing and digesting research achievements of domestic and abroad scholars about nonlinear optimization technique, the paper puts forward several improved measures as following:1) Individual of parents participates the competition of filial generation in the operation of Cross and Aberrance. It is to say that using the strategy of local preserving good individual ensures the astringency of the algorithm.2) The paper uses the parameter of the model bit's variance to adjust the probability of aberrance. It improves the performance of the algorithm by adjusting the probability of aberrance. If we use the measure along with the strategy of local preserving good individual, as long as the target function is choosed reasonably, then the improved algorithm could avoid prematurity phenomenan and has a steady constringency. The algorithm could constring to the best uncoil or the approximative best uncoil after a certain era. It adapts to the discrete problems more.3) Preserving the best individual, the global best uncoil would be obtained as soon as possible, which can test the astringency of the algorithm ulteriorly.4) For overcoming cheated problem of genetic algorithm, the paper introduces Boltzmann existent mechanism in genetic operating so that bad result of genetic algorithm also can be accepted, which keeps variety of the colony to avoid the loss of the best uncoil.5) The paper sums up diverse conversion system, and puts forward a all-purpose algorithm that adapts to from binary system to decimal system, Which may computes the same example through using different conversion system and different precision, till the satisfactory uncoil is obtained. This paper tris to find the coding mechanism that gives attention to the short length of chromosome ( It needs smaller memory space) and the high precision in seismic inversion.Through a great deal of test and analysis, we sure that the improved GA is effective and practical.